J Oral Maxillofac Surg 65:645-650, 2007 Evaluation of Sedation Failure in the Outpatient Oral and Maxillofacial Surgery Clinic Figen Cizmeci Senel, DDS, PhD,* James M. Buchanan, Jr, DDS,† Ahmet Can Senel, MD,‡ and George Obeid, DDS§

Purpose: Our goal was to report on the incidence of sedation failures in our outpatient oral surgery clinic. Sedation failure is the inability to complete a procedure under intravenous sedation. There is very little in the oral surgery literature on this subject. Materials and Methods: Proper Institutional Review Board approval was obtained from the appro- priate governing body for this project. The medical records of 539 intravenous sedation patients treated at the Oral and Maxillofacial Surgery Clinic at our institution were retrospectively evaluated to determine the incidence of failed sedation. Patients sedated with midazolam and fentanyl were placed in group A. There were 323 patients in group A. We placed patients sedated with midazolam, fentanyl and methohexital into group B. There were 216 patients in group B. The gender, medical history, type of procedure being performed, amount of drug given, and the patient’s vital signs throughout the proce- dure were recorded. Results: There were 9 failed sedations with a rate of 1.6% (9/539); 3 in group B (1%) and 6 in group A (2%). Five of our failures were undergoing multiple tooth extractions. Two of the failures were undergoing surgical removal of impacted third molars. Two patients underwent mandibular fracture reduction. Failure was attributed to increased agitation and combativeness, uncontrolled hypertension, tachychardia and desaturation. Conclusion: The mandible fracture population and multiple teeth extraction patients had higher rates of failure than other groups. This may be the result of procedure length, type of procedure, or a preoperative anxiety and attitude toward treatment expressed by the patient making sedation unpre- dictable. Level of training and experience of the practitioner may contribute to sedation failure. These results allow us to develop a prospective study protocol of outpatient sedation and to quantify more detailed information about preoperative anxiety, medical status, and social history than we had available during our chart review. More specific conclusions may help us determine if certain patient populations are at a higher risk for failed sedations. © 2007 American Association of Oral and Maxillofacial Surgeons J Oral Maxillofac Surg 65:645-650, 2007

Deep sedation/general anesthesia (DS/GA) is often the inability to continually maintain an airway inde- indicated for the relief of anxiety for outpatient oral pendently or respond purposefully to physical stimu- surgery. DS/GA is defined as an induced state of de- lation or verbal command, and is produced by a pressed consciousness or unconsciousness accompa- pharmacologic or nonpharmacologic method or com- nied by partial loss of protective reflexes, including bination thereof.1,2 In combination with local anes-

*Formerly, Research Fellow, Department of Oral and Maxillofa- §Chairman, Department of Oral and Maxillofacial Surgery, Wash- cial Surgery, Washington Hospital Center, Washington, DC; Cur- ington Hospital Center, Washington, DC. rently, Assistant Professor, Department of Oral and Maxillofacial Address correspondence and reprint requests to Dr Cizmeci Surgery, Karadeniz Technical University, Faculty of Dentistry, Senel: Department of Oral and Maxillofacial Surgery, Faculty of Trabzon, Turkey. Dentistry, Karadeniz Technical University, 61080 Trabzon, Turkey; †Former Senior Resident, Department of Oral and Maxillofacial e-mail: [email protected] Surgery, Washington Hospital Center, Washington, DC. © 2007 American Association of Oral and Maxillofacial Surgeons ‡Associate Professor, Department of Anesthesiology and Reani- 0278-2391/07/6504-0009$32.00/0 mation, Karadeniz Technical University, Faculty of Medicine, doi:10.1016/j.joms.2006.06.252 Trabzon, Turkey.

645 646 SEDATION FAILURE IN THE OUTPATIENT OMS CLINIC thesia, it is a safe and effective method of treatment. Table 1. SURGICAL PROCEDURES PERFORMED However, it is not always effective in allowing the physician to complete the planned oral surgical pro- Total cedure. On occasion, a procedure is left unfinished Procedures Patients due to patient combativeness and discomfort in spite Third molar removal 265 of increases in sedative doses. Multiple tooth extraction 157 In a review by Egelhoff et al3 of 6,006 patients Single extraction 31 sedated for diagnostic and invasive procedures, a 1% Implant placement 21 sedation failure rate utilizing different drug regimens Closed reduction of mandible fracture 17 including pentobarbital, chloral hydrate, and other Arch bar removal 13 4 regimens was found. Hoffman et al reviewed multi- Bone graft 8 ple complications encountered during non-anesthesi- Sinus lift 6 ologist-administered sedation and found a 1.35% rate Biopsy 5 of sedation failure out of 960 cases of radiologic pro- Canine exposure 5 Arthrocentesis 4 cedures, emergency department procedures, and in- Excision of cyst 2 5 vasive medical interventions. Slovis et al found a Torus excision 1 failure rate of 1.4% (40 of 2,857). This was a study of Ridge expansion 1 pediatric patients sedated for imaging procedures uti- Alveoplasty 1 lizing a drug regimen of pentobarbital, chloral hy- Soft tissue reconstruction 1 Removal of sialolith from drate, midazolam, diazepam and fentanyl. submandibular duct 1 Mason et al6 looked at sedation failure rates for magnetic resonance imaging and computed tomogra- Senel et al. Sedation Failure in the Outpatient OMS Clinic. J Oral Maxillofac Surg 2007. phy studies of pediatric patients using different drug regimens. He found a 0.6% sedation failure rate using pentobarbital and midazolam; a 1% sedation failure istics of the neck to rule out tracheal deviation, rate using pentobarbital, midazolam and fentanyl; and masses, or other mechanical interferences that would a 3% sedation failure rate using pentobarbital alone. compromise the patient’s ability to breathe under There is very little in the oral and maxillofacial DS/GA. The patients are given clear instructions to surgery literature investigating DS/GA failures. The have nothing to eat or drink after midnight prior to invasive procedures referred to in the above studies their appointment except for their medications; they included central line placement, laceration closure, can be taken with a sip of water. and other less stressful procedures. Our purpose in In group A, the procedures were started with 0.03 to this study was to determine the incidence of failed 0.06 mg/kg midazolam and 1.5 ␮g/kg fentanyl. Two DS/GA in patients undergoing oral and maxillofacial percent lidocaine with 1:100,000 epinephrine was ad- surgery and identify risk factors for failed sedation. ministered to all patients for local anesthesia. According to patients’ responses to local anesthesia administration, body habitus, etc, an additional dose of midazolam in Study Design the range of 0.03 to 0.15 mg/kg was added and then the The medical records of 539 patients treated under procedure was started. It was uncommon to administer DS/GA at the Oral and Maxillofacial Surgery outpa- more than 1.5 ␮g/kg of fentanyl. tient clinic at our institution from December 2000 In group B, the procedures began with 2 to 4 mg of through May 2002 were retrospectively reviewed to midazolam and 100 ␮g fentanyl. A bolus dose of determine the incidence of failures. We separated our methohexital to desired levels of anesthesia in doses patient population into group A and group B. Group of 10 to 15 mg was administered. The total dose of A patients were sedated using intravenous (iv) fenta- methohexital varies from 50 to 200 mg, depending on nyl and midazolam. Group B patients received fenta- patient body habitus, length of procedure, and overall nyl, midazolam, and methohexital. There were 323 response to anesthesia and the procedure being per- patients treated in group A (129 male and 194 female) formed. We used lidocaine 2% with 1:100,000 epi- aged between 12 and 80 years, with a mean age of 46 nephrine to achieve local anesthesia. No patients years. Group B had a total of 216 patients (87 male were converted from group A to group B by adding and 129 female) with an age range between 13 and 55 methohexital to complete a DS/GA. The procedures years, with a mean age of 35 years. were performed by a senior resident who was super- All patients included ASA I–II–III and had no con- vised by a member of the attending staff. traindications to the study medications. Appropriate A summary of the surgical procedures performed is medical consults are obtained when necessary. We shown in Table 1. We defined failed sedation as the also examine the oral airway and physical character- inability to complete the planned procedure. SENEL ET AL 647

Results tion about dental procedures should be assessed when studying patient populations that may fail. A Nine failed DS/GA were found in all patients (1.6%). preoperative dental anxiety level can be evaluated Three of the 9 were in group B (1%) and the other 6 using the Corah Dental Anxiety Scale. This was de- were in group A (2%). DS/GA failures were due to scribed for use in general dentistry and later modified increased levels of agitation and combativeness by the for oral surgical procedures. This is a questionnaire of patient, blood pressure above the level safe to con- the patient’s attitude to certain dental scenarios. A tinue the procedure, paroxysmal ventricular contrac- numerical score is given, based on how the questions tions manifested during the sedation, and paroxysmal are answered, that correlates to a high or low level of ventricular contractions with desaturation. Patients anxiety by the patient toward their oral surgical treat- who failed DS/GA ranged in age from 15 to 54 years ment.13 of age, with a mean age of 28 years. Our patients choose to have their procedures per- Five of 157 (3.18%) patients having multiple teeth formed under DS because they have an increased extracted experienced DS/GA failure. Two of 265 level of anxiety toward dental procedures in general. patients (0.75%) undergoing surgical removal of im- It has been shown that DS versus local anesthetic pacted third molars and 2 of 17 patients (11.76%) decreases this anxiety level and allows the oral undergoing mandibular fracture reduction experi- and maxillofacial surgeon to perform surgery enced DS/GA failure. successfully.13 Malviya et al7 found that patients who A summary of the failures is shown in Table 2. failed noninvasive radiologic procedure sedation were less well adapted to medical treatment than those who were successful. They conclude that a Discussion preoperative assessment of the patient’s tempera- Nine of 539 patients failed DS with a rate of 1.6%. ment should be obtained to determine whether pa- This is consistent with failure rates reported by oth- tients can be successfully sedated. There is nothing ers. However, it is hard to compare our study with published that studies anxiety levels on the success or those of others. Most studies used medication regi- failure of sedation in adult populations. mens different from ours, including diazepam, pento- The alleviation of preoperative dental fear can be barbital, meperidine, chloral hydrate, and propofol. facilitated by a comforting atmosphere in the opera- Oral and maxillofacial procedures are also always tory. This soothing environment can be very difficult invasive in nature, causing pain and discomfort to the to establish in a hospital oral surgery clinic for many patient. Radiological procedures are noninvasive even reasons. Because of this, a patient’s sense of security though some may cause claustrophobia and anxiety in may not be established as effectively as in a private pediatric populations.7,8 Oral surgery causes pain and office. It is difficult to objectively evaluate your clinic requires local anesthesia and iv medications when environment, but we need to consider the effect of indicated.9-11 Most studies looked at sedation failures clinic atmosphere on our patient’s perception of their in noninvasive or minimally invasive procedures. treatment when evaluating sedation failure. They considered central line placement, suturing of Oral surgery often uses local anesthesia in addition lacerations, and starting an iv as invasive. The study by to the iv medications used. Inadequate local anesthe- Hoffman et al4 included noninvasive and invasive pro- sia creates an obvious challenge during DS and cannot cedures performed in a pediatric emergency depart- generally be overcome by an increase in the dose of a ment and found a 1.35% (13/960) sedation failure sedative agent. Some of our failures may be due to rate. Pena and Krauss,12 in a similar study of invasive missed mandibular blocks and a lack of profound and noninvasive procedures in a pediatric emergency anesthesia. Successful anesthesia of an abscessed department, found a 0.85% (10/1180) failure rate. tooth with swollen surroundings can, at times, be Invasive procedures performed in the emergency de- difficult. Some procedures can be quite uncomfort- partment are often shorter and less stressful than able even with good local anesthesia. Two of 17 those performed by an oral surgeon. Stress from pain, (11.76%) of our mandibular fracture group failed DS. pressure and auditory stimuli of a roto-osteotome dur- The manipulation of a fractured mandible can be ing oral surgery are very challenging to manage dur- quite uncomfortable. Obtaining adequate local anes- ing DS. thesia to cover a fractured site as well as the surround- The above studies include pediatric and adolescent ing bruised area is difficult to accomplish. The painful populations, whereas our study includes older pa- stimuli in an already suffering and frightened patient tients. The maturity of a patient may affect their level are enough to arouse them during DS. of anxiety toward oral surgery, which can affect the It is reasonable to expect that longer DS/GA may be success of a DS. Even though our DS failures were more prone to failure, but this was not found to be a mostly adults (aged 15-54 years), a preconceived no- significant factor in our study. Eight of our failures 648

Table 2. SUMMARY OF SEDATION FAILURES

Procedure Cancelled Local Vital Signs Cause of Age Relevant Type of After Fentanyl Midazolam Methohexitone Anesthesia During Patient Group Failure (yrs) Gender Medical History Surgery (minutes) (mcg) (mg) (mg) (mg) Procedure 2 A Agitated and 18 M None Closed 30 100 4 0 120 Within combative reduction normal fracture mandible 4 A Agitated and 21 M Frequent Closed 20 100 5 0 150 Within combative alcohol reduction normal consumption fracture mandible 5 A Significant 54 F HTN and an Multiple 10 100 4 0 150 170/110 hypertension old stroke extractions mmHg 6 A Agitated and 24 F Mild anemia Third 20 200 5 0 120 Within combative and molar normal hypothyroidism surgery 8 A Significant 47 M HTN/diabetes/ Multiple 25 0 4 0 120 220/120 hypertension impaired tooth mmHg renal extraction function CLINIC OMS OUTPATIENT THE IN FAILURE SEDATION 9 A Agitated and 28 M None Full mouth 75 200 6 0 Within combative extractions normal 1 B PVC 25 F None Third 30 100 4 40 120 Within molar normal surgery 3 B Tachycardia 15 F Blind Multiple 10 100 3 20 120 Pulse and extractions 125– desaturation 160 beats/ min 7 B Agitated and 21 M HIV Multiple 10 100 6 100 170 Within combative tooth normal extraction Senel et al. Sedation Failure in the Outpatient OMS Clinic. J Oral Maxillofac Surg 2007. SENEL ET AL 649 were procedures terminated at less than 30 minutes. other patient had no history of hypertension but de- Patient 9 became agitated and combative after 75 veloped premature ventricular contractions (PVCs) minutes; 19 of 26 teeth were removed prior to termi- after initiating the procedure. nating the procedure due to increased levels of agita- The last patient had a history of hypertension and tion and combativeness. He received 6 mg midazolam kidney failure and was on dialysis. These patients and 200 ucg of fentanyl. He was a 28-year-old healthy commonly develop hyper- and hypotension during African American male, but was combative and anx- sedation. Their volume status controlled by dialysis ious from the start of the procedure, making the plays a large role in how they respond to sedative extraction of his teeth quite challenging. As in all the medications and painful stimuli. All of these patients other procedures, the operator was a resident who had a thorough history and physical and were consid- may not have had enough experience to manage a full ered safe to undergo DS in the oral surgery clinic. mouth of extractions on a difficult patient. This may Physiologic variations outside of the norm occur dur- have also contributed to the prolongation of the pro- ing DS. Epinephrine found in local anesthesia and cedure. released by the body during pain and stress causes Five failures were due to patient agitation and com- tachychardia, increased mean arterial pressure, in- bativeness. They only received 4 to 6 mg of versed creased cardiac output, increased systemic vascular and 100 to 200 ucg of fentanyl. This is a significant resistance, and cardiac dysrythmias, most commonly dose, and with adequate local anesthesia should allow manifested as PVCs. Methohexital has excitatory ef- for successful surgery. It is important to consider fects, including tachychardia.14,15 whether increased doses of versed with added local The DS discontinued because of physiologic param- anesthesia would have allowed us to complete the eters outside of the norm were reasonable decisions. procedure. However, with the increased level of alert- Hypertension often results from painful stimuli, but ness and the combativeness, it was believed that we could not control it with increased anesthesia. It is deepening the sedation would not allow us to accom- our professional judgment to discontinue a procedure plish the remaining invasive part of the procedure. after administering iv hypotensive medications or li- The patient’s confidence in the process was already docaine to control PVCs. We believe that we were lost, and it would be very hard to calm their fear once neither too cautious nor too aggressive in completing they reach this distrustful phase. One would think the procedures, but this is quite subjective and should that converting patients from group A to group B be considered when evaluating DS failure. There is would solve this problem by adding methohexital to speculation that alcohol abuse leads to an up-regula- the regimen, but this was not attempted with our tion of the gamma aminobutyric acid A (GABAA)re- patients. There is evidence that paradoxical reactions ceptor in certain populations. This up-regulation to barbiturates and benzodiazepines occur in select causes a tolerance to gabaminergic drugs. Cheng et populations. These reactions are manifested as overt al16 studied sedation failure using versed in patients excitement in the presence of pain with the adminis- being treated for chronic pain related to cancer in the tration of barbiturates. This is explained by an overall intensive care unit setting. They believed that an up- suppression of inhibitory mechanisms and a lack of regulation of the GABAA receptor is a possible cause sedative effect by the medication. Benzodiazepines of the ineffectiveness of benzodiazepines. The GABAA have documented paradoxical reactions including receptor is the effector center for midazolam and nightmares, combativeness and other untoward ef- methohexital. It is unknown how to determine which 14,15 fects. Morphine and other narcotics can cause patients are susceptible to this effect on the GABAA histamine release. This has not been seen with fenta- receptor, but we can guess that midazolam may be nyl. Three procedures in our review were due to unsuccessful because of the physiologic effects of cardiovascular complications developing after the ini- chronic alcohol use on the central nervous system tiation of the procedure. General anesthetics alter a resulting in an increased tolerance to midazolam. We patient’s response to painful stimuli and their internal can infer a clinically relevant cause-and-effect relation- hemodynamic control mechanisms. Hoffman et al’s ship in the population of chronic alcohol users, but study4 states that non-anesthesiologist-administered many of these patients are successfully treated with sedation is safe and effective, but that complications DS. More research is needed in pharmacology to are avoided by adhering to a strict presedation proto- show a clear effect of alcohol on the physiologic and col, including a risk assessment based on factors such pharmacologic function of the GABAA receptor in as medical conditions and patient history of anes- chronic alcohol users. thetic complications. Pre-existing medical conditions The purpose of our survey was to evaluate DS that are well controlled should not exclude one as a failures in our clinic. Our mandible fracture popula- candidate for DS. One of these 3 failures was a patient tion and population of patients with multiple teeth with a history of well-controlled hypertension. The extractions had higher rates of failure than other 650 SEDATION FAILURE IN THE OUTPATIENT OMS CLINIC groups. This may be due to the length of the DS, type 5. Slovis TL, Parks C, Reneau D, et al: Pediatric sedation: Short- term effects. Pediatr Radiol 23:345, 1993 of procedure, level of stimuli, or a preoperative anx- 6. Mason KP, Zurakowski D, Karian VE, et al: Sedatives used in iety and attitude toward treatment expressed by the pediatric imaging: Comparison of IV pentobarbital with pento- patient that makes DS more difficult. This study al- barbital with midazolam added. AJR Am J Roentgenol 177:427, 2001 lows us to develop a protocol for a prospective study 7. Malviya S, Voepel-Lewis T, Eldevik OP, et al: Sedation and of outpatient DS procedures and to quantify more general anesthesia in children undergoing MRI and CT: Adverse detailed information about patients’ preoperative anx- events and outcomes. Br J Anaesth 84:743, 2000 8. Mason KP, Koka BV: New challenges in anesthesia: New prac- iety, medical status and social history than we had tice opportunities/anesthesia for pediatric radiology. Anesthe- available during our chart review. More specific con- siol Clin North America 2:17, 1999 clusions may help the oral surgeon determine if cer- 9. Moore PA, Finder RL, Jackson DL: Multidrug intravenous seda- tion/determinants of sedative dose of midazolam. Oral Surg tain patient populations are at a higher risk for failed Oral Med Oral Pathol 84:5, 1997 DS. 10. Ochs MW, Tucker MR, Owsley TG, et al: The effectiveness of flumazenil in reversing the sedation and amnesia produced by intravenous midazolam. J Oral Maxillofac Surg 48:240, 1990 11. Parworth LP, Frost DE, Zuniga JR, et al: Propofol and fentanyl References compared with midazolam and fentanyl during third molar surgery. J Oral Maxilofac Surg 56:447, 1998 1. Coyle TT, Helfrick JF, Gonzalez ML, et al: Office-based ambu- 12. Pena BM, Krauss B: Adverse events of procedural sedation and latory anesthesia: Factors that influence patient satisfaction or analgesia in a pediatric emergency department. Ann Emerg dissatisfaction with deep sedation/general anesthesia. J Oral Med 34:483, 1999 Maxillofac Surg 63:163, 2005 13. Bell GW, Kelly PJ: A study of anxiety and midazolam-induced 2. Perrott DH, Juen JP, Andresen RV, et al: Office-based ambula- amnesia in patients having lower third molar teeth extracted. tory anesthesia: Outcomes of clinical practice of oral and max- Br J Oral Maxillofac Surg 38:596, 2000 illofacial surgeons. J Oral Maxillofac Surg 61:983, 2003 14. Hardman J, Limbird L, Molinoff P, et al (eds): Goodman & 3. Egelhoff JC, Ball WS Jr, Koch BL, et al: Safety and efficacy of Gilman’s The Pharmacological Basis of Therapeutics. New sedation in children using a structured sedation program. AJR York, NY, McGraw-Hill, 1996 Am J Roentgenol 168:1259, 1997 15. Reves JG, Glass PSA, Lubarsky DA, et al: ltravenous nonopioid 4. Hoffman GM, Nowakowski R, Troshynski TJ, et al: Risk reduc- anesthetics, in Miller R (ed): Miller’s Anesthesia. New York, tion in pediatric procedural sedation by application of an Amer- NY, Churchill Livingstone, 2004, pp 317-437 ican Academy of Pediatrics/American Society of Anesthesiolo- 16. Cheng C, Roemer-Becuwe C, Pereira J: When midazolam fails. gists process model. Pediatrics 109:236, 2002 J Pain Symptom Manage 23:256, 2002